A method for rapid assessment of seismic vulnerability of wind farm based on physical seismic source simulation

Abstract

The present application relates to a kind of wind farm seismic vulnerability fast evaluation method based on physical seismic source simulation, comprising the following steps: step S1: obtaining the parameter of active fault near target wind farm, establish regional one-dimensional crustal velocity structure model;Step S2: construct kinematic hybrid seismic source model, in wave number domain in combination low wave number certainty and high wave number randomness slip distribution, describe fault rupture space-time evolution;Step S3: using propagation matrix method in frequency-wave number domain solves layered half-space dislocation point source Green function, by finite fault sub-source superposition synthesis multiple point, multiple component broadband ground motion time history;Step S4: establish wind power tower nonlinear finite element model, input synthesis ground motion and carry out dynamic time history analysis;Step S5: define damage state, using cloud chart method establishes probabilistic seismic demand model, fitting vulnerability curve, assesses the seismic risk of wind farm under different fault rupture scenarios.The present application realizes the whole process physical simulation and fast vulnerability evaluation from seismic source to structure, provides scientific basis for wind farm seismic design and site selection decision.

Description

Technical Field

[0001] This invention belongs to the field of wind power structure seismic resistance technology, and in particular, it is a rapid assessment method for the seismic vulnerability of wind farms based on physical seismic source simulation. Background Technology

[0002] Wind power, as a clean and renewable energy source, has developed rapidly in my country in recent years, with its installed capacity continuously expanding. As wind power development extends to seismically active areas such as the west and coastal regions, more and more wind farms are being built near active fault zones, highlighting the increasing seismic risks they face. Wind turbine towers typically consist of a conical steel tower, blades, nacelle, and foundation. They are slender, tall, and flexible structures with a large concentrated mass at the top, a long fundamental period, and a low damping ratio. Under seismic loading, they are prone to dynamic amplification and resonance effects, and may exhibit complex failure modes such as tower yielding and buckling. Therefore, accurately assessing the vulnerability of wind farms under potential seismic loading is crucial for their seismic design, site selection decisions, and risk management.

[0003] The accuracy of seismic vulnerability assessment depends on the rationality of the seismic motion input. Source characteristics, propagation paths, and site conditions are key factors influencing the seismic response of wind turbine towers. However, most existing methods for assessing the seismic vulnerability of wind turbine towers rely on historical seismic motion records from different regions, approximating the target site's response spectrum through spectral matching or amplitude adjustment before using it as input. These methods have inherent limitations: firstly, historical records are scattered, making it difficult to reflect the true rupture characteristics of specific active faults near the target wind farm, such as fault geometry, slip distribution, and rupture directionality; secondly, for large-scale wind farms, it is usually assumed that all turbine locations within the farm experience uniform seismic excitation, neglecting the spatial variability of seismic motion caused by differences in source mechanisms and wave propagation paths.

[0004] Furthermore, near-fault ground motions exhibit significant directional effects, and their spectral composition changes with epicentral distance. As the propagation distance increases, the high-frequency components of the ground motion gradually attenuate, while the long-period components relatively increase. The long-period energy poses a serious threat to the highly flexible wind turbine tower structure, which has a relatively long fundamental period. These physical phenomena are difficult to reproduce using traditional methods based on historical records. Therefore, there is an urgent need to develop a vulnerability assessment method that comprehensively considers the physical mechanisms of the earthquake source, propagation path, and the entire site process, and reflects the spatial differences in ground motions within the wind farm, in order to more realistically and accurately assess the disaster risk posed by specific active faults to wind farms. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing methods that do not consider the physical rupture process of specific faults and the spatial variability of ground ground motion, and to propose a rapid assessment method for the seismic vulnerability of wind farms based on physical source simulation.

[0006] The technical solution adopted in this invention is: Step S1: Obtain the geometric and kinematic parameters of the active fault in the area where the target wind farm is located, and establish a one-dimensional crustal velocity structure model of the region by combining exploration data or global crustal models. This model includes the thickness of each layer, P-wave velocity, S-wave velocity, density, and quality factor (parameters describing the energy attenuation characteristics of seismic waves during propagation in the medium). The geometric and kinematic parameters of the active fault include at least the strike, dip angle, slip angle, length, width, and burial depth. Step S2: Construct a kinematic hybrid source model based on the parameters in Step S1. Determine the seismic moment and average slip according to the global parameters. Use the source parameter calibration law (i.e., the statistical law describing the empirical relationship between parameters such as fault size, seismic moment and concave / convex body area) to determine the local source parameters. In the wavenumber domain, take the corner wavenumber (i.e. the transition wavenumber between deterministic and random slip distribution) as the boundary to smoothly combine low wavenumber deterministic slip and high wavenumber random slip. Obtain the spatial domain hybrid slip distribution through inverse Fourier transform, and determine the rupture time and rise time to generate a complete source model. Step S3: Solve for the Green's function of the layered half-space dislocation point source in the frequency-wavenumber domain using the propagation matrix method. Discretize the finite fault into multiple sub-sources, and synthesize the three-component broadband ground motion time history of multiple virtual observation points by convolving and superimposing the Green's function of each sub-source with the source time function and the seismic moment. Step S4: Establish a nonlinear finite element model of the wind turbine tower, use the synthetic ground motion as input for dynamic time history analysis, and obtain the structural response; Step S5: Define multi-level damage states, establish a probabilistic earthquake demand model using the cloud map method (i.e., directly use the unmodulated ground motion for time history analysis, and perform linear regression of the ground motion intensity index and engineering demand parameters in the logarithmic space), and combine the log-normal distribution to fit the vulnerability curve to assess the earthquake risk of wind farms under different fault rupture scenarios. Furthermore, in step S2, the wavenumber domain sliding distribution function of the hybrid source model... as follows: ; In the formula, For a deterministic sliding distribution, It is a random sliding distribution. and These are the wavenumbers along the strike and dip directions, respectively. To combine the functions, control the transition between the deterministic and random components at the corner wavenumber.

[0007] Furthermore, in step S3, the ground motion at observation point x Obtained through the superposition of finite fault sub-sources: ; In the formula, For the first A rectangular sub-source at the observation point The displacement Green's function at the location; STF is the source time function. The rise time of this sub-source. This is the initial rupture time of the subsource; This represents the scalar seismic moment of the sub-source; the symbol * denotes temporal convolution operation; This represents the summation of all rectangular sub-sources i and j on the fault plane, where i and j are the sub-source numbers along the fault strike and dip, respectively.

[0008] The Green's function is solved using the propagation matrix method: the elastic wave equation is subjected to Fourier-Hankel transform to the frequency-wavenumber domain, and the displacement-stress relationship of each layer is recursively derived using the layer propagation matrix. The displacement field is solved by combining the boundary conditions, and then the Green's function in the time-space domain is obtained by inverse transformation.

[0009] Furthermore, in step S5, the formula for the probabilistic earthquake demand model is as follows: ; In the formula, For engineering requirements, take the tower top displacement, acceleration, or section stress. IM The peak ground acceleration is used as an indicator of seismic intensity. a , b These are the regression coefficients; The vulnerability function is given by formula (4): ; In the formula, Indicator of local seismic intensity Values At that time, structural seismic demand Exceeding structural capacity That is, the probability of reaching or exceeding the corresponding damage state; The cumulative distribution function of the standard normal distribution; This represents the median structural capacity corresponding to this damage state. This represents the median of the corresponding earthquake demand. The logarithmic standard deviation of the earthquake demand parameter. Due to the uncertainty of the structure's load-bearing capacity. Cognitive uncertainty is introduced into the modeling.

[0010] Furthermore, when the active fault mentioned in step S1 is a long fault zone with segmented independent ruptures, the fault zone is divided into several sub-fault segments; according to steps S2 to S5, source models are constructed for the independent rupture scenarios of each sub-fault segment, the combined ground motion is calculated, and the structural response and vulnerability curves are calculated; by comparing the vulnerability curves under the rupture scenarios of each sub-fault segment, the sub-fault segment with the highest damage exceedance probability under the same ground motion intensity index or the most severe disaster consequences under the maximum possible magnitude is identified as the key fault segment for the seismic design of the dominant wind farm.

[0011] Furthermore, in step S3, multiple virtual observation stations are set up in the wind farm area according to the grid, and the ground motion time history at each station location is synthesized as the ground motion input of the corresponding wind turbine tower to characterize the spatial variability of ground motion in the wind farm; and whether to use consistent ground motion excitation is determined according to the relative deviation of the structural response of each turbine: when the maximum relative deviation of the structural response of each turbine is less than the preset threshold, consistent ground motion excitation is used to assess the vulnerability of the wind farm; otherwise, non-consistent ground motion excitation is used.

[0012] Compared with the prior art, the present invention has the following advantages and effects: This invention combines physical source simulation with structural dynamics simulation to construct a full-process modeling framework from source to structure. It employs a kinematic hybrid source model and the propagation matrix method to synthesize broadband ground motions, balancing low-frequency determinism and high-frequency randomness, thus reflecting the spatial variability of ground motions. Furthermore, it establishes a quantitative vulnerability model using a nonlinear finite element model and the contour plot method, providing a scientific basis for wind farm seismic design, site selection decisions, and insurance pricing. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating the overall implementation of the method of the present invention; Figure 2 This is a slip distribution diagram of the hybrid source model in the embodiment; Figure 3 This is a time history diagram of the ground motion acceleration synthesized in the example; Figure 4 The diagram shows the seismic vulnerability curve of the wind turbine tower in the example. Detailed Implementation

[0014] The present invention will now be clearly and completely described with reference to the accompanying drawings and specific embodiments. The described embodiments are only a part of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] This embodiment takes a wind farm adjacent to an active fault zone as the research object. Considering that the fault zone is long and has the characteristics of segmented independent rupture, it is divided into several sub-fault segments. Seismic motion simulation and structural vulnerability assessment are carried out for each sub-fault segment under the scenario of independent rupture, according to the following steps. By comparison, the key fault segments controlling the seismic design of the wind farm are identified. The overall process is as follows: Figure 1 As shown.

[0016] Step S1: Investigate the seismic geological data of the target wind farm area to determine the strike, dip angle, slip angle, length, width and burial depth of active faults; combine regional exploration data and global crustal models to establish a one-dimensional crustal velocity structure model including the thickness of each layer, P-wave velocity, S-wave velocity, density and quality factor (parameters describing the energy attenuation characteristics of seismic waves in the medium).

[0017] Step S2: Determine the seismic moment and average slip based on the global fault parameters obtained in Step S1, and determine the local source parameters such as the concave-convex body (i.e., the concentrated slip area on the fault plane where the slip is significantly higher than the average value) based on the source parameter calibration law (i.e., the statistical law describing the empirical relationship between parameters such as fault size, seismic moment and concave-convex body area), and construct a kinematic hybrid source model. Wavenumber domain slip distribution function of hybrid source model for: (1); In the formula, For a deterministic sliding distribution, It is a random sliding distribution. and These are the wavenumbers along the strike and dip directions, respectively. To combine the functions, control the transition between the deterministic and random components at the corner wavenumber.

[0018] In the wavenumber domain, the low-wavenumber deterministic slip and high-wavenumber random slip are smoothly combined according to formula (1). The spatial domain hybrid slip distribution is obtained through inverse Fourier transform, and the rupture time and rise time of each sub-source are determined to generate a complete kinematic hybrid source model. The resulting slip distribution is as follows: Figure 2 As shown.

[0019] Step S3: Use the propagation matrix method to solve the dislocation source Green function in the frequency-wavenumber domain under the velocity structure of step S1; discretize the finite fault in step S2 into multiple rectangular sub-sources, and convolve and superimpose the Green function of each sub-source with its source time function and seismic moment according to formula (2) to synthesize the three-component broadband ground motion time history of multiple virtual observation points in the wind farm area.

[0020] Earthquake at observation point x The formula for superimposing finite fault sources is: (2); In the formula, For the first A rectangular sub-source at the observation point The displacement Green's function at the location; STF is the source time function. The rise time of this sub-source. This is the initial rupture time of the subsource; This represents the scalar seismic moment of the sub-source; the symbol * denotes temporal convolution operation; This represents the summation of all rectangular sub-sources i and j on the fault plane, where i and j are the sub-source numbers along the fault strike and dip, respectively.

[0021] The Green's function is solved using the propagation matrix method: the elastic wave equation is subjected to Fourier-Hankel transform to the frequency-wavenumber domain, and the layer propagation matrix is ​​used to recursively deduce the relationship between displacement and stress of each layer. The displacement field is solved by combining the boundary conditions, and then the Green's function in the time-space domain is obtained by inverse transformation.

[0022] In this embodiment, multiple virtual stations are deployed in a regular grid within the wind farm area. Corresponding ground motions are synthesized for each sub-fault segment experiencing independent rupture. The synthesis results are as follows: Figure 3 As shown.

[0023] Step S4: Establish a three-dimensional nonlinear finite element model of the wind turbine tower in the finite element software. The tower body adopts nonlinear beam-column elements with fiber sections arranged along the height (i.e., the component section is discretized into multiple fibers and assigned material constitutive relations respectively). The blades and nacelle are equivalent to eccentric concentrated mass points at the top of the tower. Input the ground motion time history synthesized in Step S3 simultaneously in three directions to perform nonlinear dynamic time history analysis and extract structural responses such as tower section stress, tower top displacement and acceleration. Perform the above analysis on the wind turbines at each virtual station corresponding to the turbine location and compare the relative deviation of the structural response at each turbine location: when the maximum relative deviation is less than the preset threshold, the vulnerability of the wind farm can be assessed by uniform ground motion excitation throughout the entire field; otherwise, non-uniform ground motion excitation is used to balance calculation accuracy and calculation cost.

[0024] Step S5: Define the multi-level damage states of the key subsystems of the tower and nacelle; for the rupture scenarios of each sub-fault segment, uniformly sample the magnitude to generate a large number of earthquake scenarios and synthesize the corresponding ground motions, and calculate the structural response in batches according to step S4; perform logarithmic linear regression on the ground motion intensity index and structural requirement parameters according to formula (3), and then fit the vulnerability curves of each damage state according to formula (4), such as Figure 4 As shown.

[0025] The probabilistic earthquake demand model is as follows: (3); In the formula, For engineering requirements, parameters such as tower top displacement, acceleration, or section stress can be used. For seismic intensity indicators, peak ground acceleration can be used. a , b is the regression coefficient.

[0026] The vulnerability function is: (4); In the formula, Indicator of local seismic intensity Values At that time, structural seismic demand Exceeding structural capacity That is, the probability of reaching or exceeding the corresponding damage state; The cumulative distribution function of the standard normal distribution; This represents the median structural capacity corresponding to this damage state. This represents the median of the corresponding earthquake demand. The logarithmic standard deviation of the earthquake demand parameter. Due to the uncertainty of the structure's load-bearing capacity. Cognitive uncertainty is introduced into the modeling.

[0027] By comparing the vulnerability curves under rupture scenarios of each sub-fault segment, the key fault segments that dominate the seismic design of wind turbine towers in terms of ground motion intensity and structural response can be identified, thus providing a basis for wind farm site selection and seismic fortification.

Claims

1. A rapid assessment method for the seismic vulnerability of wind farms based on physical seismic source simulation, characterized in that, Includes the following steps: Step S1: Obtain the geometric and kinematic parameters of the active fault in the area where the target wind farm is located, and establish a one-dimensional crustal velocity structure model of the region by combining exploration data or global crustal models, including the thickness of each layer, P-wave velocity, S-wave velocity, density and quality factor; the geometric and kinematic parameters of the active fault include at least strike, dip angle, slip angle, length, width and burial depth. Step S2: Based on the fault parameters obtained in Step S1, construct a kinematic hybrid source model; the model smoothly combines the low wavenumber deterministic slip distribution and the high wavenumber random slip distribution in the wavenumber domain with the corner wavenumber as the boundary, and obtains the spatial domain hybrid slip distribution through inverse Fourier transform, and determines the rupture time and rise time of each sub-source, generating a complete source model describing the spatiotemporal evolution of fault rupture; Step S3: Use the propagation matrix method to solve the Green's function of the dislocation point source in the layered crustal half space in the frequency-wavenumber domain; discretize the finite fault model in step S2 into multiple rectangular sub-sources, and convolve and superimpose the Green's function of each sub-source with its source time function and scalar seismic moment to synthesize the three-component broadband ground motion time history of multiple virtual observation points in the wind farm area. Step S4: Establish a nonlinear finite element model of the wind turbine tower, and use the ground motion time histories of each turbine location synthesized in Step S3 as input to perform nonlinear dynamic time history analysis and obtain the structural response; Step S5: Define the multi-level damage state of the key subsystems of the wind turbine tower. Based on the ground motion intensity index and structural requirement parameters obtained in Step S4, establish a probabilistic earthquake demand model using the cloud map method. Combine the log-normal cumulative distribution function to generate vulnerability curves and assess the earthquake risk of the wind farm under different fault rupture scenarios.

2. The method according to claim 1, characterized in that, In step S2, the wavenumber domain sliding distribution function of the hybrid source model for: ； In the formula, For a deterministic sliding distribution, It is a random sliding distribution. and These are the wavenumbers along the strike and dip directions, respectively. To combine the functions, control the transition between the deterministic and random components at the corner wavenumber.

3. The method according to claim 1, characterized in that, In step S3, the propagation matrix method solves for the Green's function in the following way: The elastic wave equation is subjected to a Fourier-Hankel transform to the frequency-wavenumber domain; the layer propagation matrix is ​​used to recursively establish the displacement-stress relationship of each layer; the displacement field excited by the dislocation source is solved by combining the free surface and half-space radiation conditions; then, the inverse Fourier-Hankel transform is used to obtain the Green's function in the time-space domain; the seismic motion at observation point x... Obtained through sub-source superposition: ； In the formula, For the first A rectangular sub-source at the observation point The displacement Green's function at the location; the STF is the source time function, describing the time process of seismic energy release from each sub-source. The rise time of this sub-source. This is the initial rupture time of the subsource; This is the scalar seismic moment of the sub-source; The symbol * represents temporal convolution operation; This represents the summation of all rectangular sub-sources i and j on the fault plane, where i and j are the sub-source numbers along the fault strike and dip, respectively.

4. The method according to claim 1, characterized in that, In step S4, the finite element model of the wind turbine tower adopts a fiber section arranged along the height, that is, the component section is discretized into multiple fibers and assigned material constitutive relations to nonlinear beam and column elements to simulate the tower. The blades and nacelles are equivalent to the eccentric concentrated mass point at the top of the tower, and the seismic load is applied simultaneously in three directions.

5. The method according to claim 1, characterized in that, In step S5, the probabilistic earthquake demand model is as follows: ； In the formula, For engineering requirements, take the tower top displacement, acceleration, or section stress. IM The peak ground acceleration is used as an indicator of seismic intensity. a , b These are the regression coefficients; The vulnerability function is: ； In the formula, Indicator of local seismic intensity Values At that time, structural seismic demand Exceeding structural capacity That is, the probability of reaching or exceeding the corresponding damage state; The cumulative distribution function of the standard normal distribution; This represents the median structural capacity corresponding to this damage state. This represents the median of the corresponding earthquake demand. The logarithmic standard deviation of the earthquake demand parameter. Due to the uncertainty of the structure's load-bearing capacity. Cognitive uncertainty is introduced into the modeling.

6. The method according to claim 1, characterized in that, When the active fault mentioned in step S1 is a long fault zone with independent segmented ruptures, the fault zone is divided into several sub-fault segments. According to steps S2 to S5, source models are constructed for the independent rupture scenarios of each sub-fault segment, the combined ground motion is calculated, and the structural response and vulnerability curves are calculated. By comparing the vulnerability curves under the rupture scenarios of each sub-fault segment, the sub-fault segment with the highest damage exceedance probability under the same ground motion intensity index or the most severe disaster consequences under the maximum possible magnitude is identified as the key fault segment for the seismic design of the dominant wind farm.

7. The method according to claim 1, characterized in that, In step S3, multiple virtual observation stations are set up in the wind farm area according to the grid, and the ground motion time history at each station location is synthesized as the ground motion input of the corresponding wind turbine tower to characterize the spatial variability of ground motion in the wind farm. The system determines whether to use consistent ground motion excitation based on the relative deviation of the structural response of each turbine location: when the maximum relative deviation of the structural response of each turbine location is less than the preset threshold, consistent ground motion excitation is used to assess the vulnerability of the wind farm; otherwise, non-consistent ground motion excitation is used.

Images